The physical properties of an oil reservoir is affected by various geologic factors. It is commonly manifested as inhomogeneity, which is usually characterized by a distinct structure on a certain spatial scale. Take for example the early Late Eogene oil deposits in the Gas Hure oil field, Qinghai Province, the paper discusses the application of the geostatistical method in analysing and predicting the spatial structural characteristics of reservoir physical properties. The spatial structural characteristics of porosity and permeability are quantitatively described with variograms. Based on the experimental variograms and fitted theoretic curves, the authors discuss the optimal interpolating prediction by using the Kriging method. The result indicates that the spatial structural distribution of porosity and permeability in the reservoir is distinct. Most of the ranges of the variograms are around 800-2000 m. The characteristics of porosity and permeability distribution in different reservoir layers are distinct from each other. This feature of porosity and permeability for different layers is mainly affected by the spatial distribution of sedimentary facies. The experimental semi variograms of porosity and permeability of each sedimentary layer can be fitted to a spherical model with the mugget effect. Therefore, the numerical values of porosity and permeability among the observation spots can be interpolated and predicted based on the model. To verify the above result, a multiple trend surface analysis model is used to predict the values of porosity and permeability.